SARMA Time Series for Microscopic Electrical Load Modeling

Comparative assessment of a quasi-sequential and a sequential approach for distribution network stochastic analysis

“…It should be noted that the 15 individuals were selected out of a set of 18 prosumers. Although a reasonable model can be obtained in most cases with a SARMA approach, a satisfactory model cannot always be defined for different considerations discussed in [2]. Additionally, the performances reached for each individual are not identical.…”

“…The two terms in (6) refer respectively to the AR and MA processes, the first term of (7) corresponds to the Seasonal AR process (SAR) of order P and the second term is the Seasonal MA process (SMA) of order Q . More information about the methodology for the definition of such microscopic models is available in [2].…”

“…Different dynamics are therefore at stake and a single model tends to average those. This remark can be made for the consumption pattern as well but for different reasons (see [2]). Sunny days lead to a quasi-deterministic bell shaped curve.…”

“…Two main approaches, respectively quasi-sequential and sequential, are compared pointing out their particularities, advantages and drawbacks using Smart Meter data. Such microscopic modeling requiring solely historical data recorded by smart meters is recent [2] and this paper intends to highlight as well the potential of this mean to build effective and efficient models. Finally, the necessity to adopt sequential models in the upcoming context is demonstrated.…”

Non‐linear hybrid approach for the scheduling of merchant underground pumped hydro energy storage

“…Practically, the hybrid tool is run for different sets of the four learning rates (by relying on a grid search). For each 4D point of the grid search, the optimisation outcomes are averaged over eight representative days (one per season, differentiated between week and weekend days), which are generated using [36] with 5 years of historical market data (from 2013 to 2017), to properly cover possible scenarios over a typical year. Based on these results, we represent (in Fig.…”